| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Removing obsolete Cluster info from repositories |
« View previous topic :: View next topic » |
| Author |
Message
|
| basevij |
 Posted: Wed Oct 02, 2002 6:23 am Post subject: Removing obsolete Cluster info from repositories |
 |
|
Newbie
Joined: 11 Jun 2002
Posts: 5
Location: Bristol, UK
|
We recently had to recreate all the Qmgrs in a cluster. This was done singly suspending clustermembership and then taking down and recreating the Qmgr. This has left us with the problem that the Cluster repositories have references to the old Qmgr instances and is causing the automatic channels to fail as they are being created using obsolete information. Is there a safe way to clear both repositories of this duff information
JB |
|
| Back to top |
|
 |
| 2189 |
| Posted: Tue Oct 22, 2002 5:23 am Post subject: |
|
|
Apprentice
Joined: 22 Oct 2002
Posts: 31
|
If this is a production system and you've got lots of qmgrs and objects then my solution might not sensible, it's also probably the last resort in any situation, however I have got this to work in a test environment.
If you can't clean out the old objects using offical MQ commands then manually clearing out the repositories shud fix the problem - pretty radical I know so don't shoot me down please.
It not that simple to get the whole thing up and running after tho. This is what I did:
1. kill the repos process for the repos managers
2. clear all messages of SYSTEM.CLUSTER.REPOSITORY.QUEUE
The repos qmgrs now have no info about any other qmgr so you need to somehow get all the qmgrs to relearn about every object. Do this as follows:
3. Alter the clusrcvr channel (e.g. the desc) on every connected qmgr. Check the clussdr chl has started, if so the repos shud now know about the qmgr. Check using DIS CLUSQMGR(*).
4. Alter every clustered object (e.g. the desc) on every qmgr in the cluster. This should send details of the clustered objects to the repos.
Hopefully all the channels will have started and there should be no cluster messages on the transmit queues of any qmgrs.
Having suggested the radical method you could always try the clustersupport pac to remove the old info first but I've never managed to get that to work very well.
_________________
2189 errors make me  |
|
| Back to top |
|
|
| RogerLacroix |
| Posted: Tue Oct 22, 2002 8:33 pm Post subject: |
|
|
Jedi Knight
Joined: 15 May 2001
Posts: 3253
Location: London, ON Canada
|
2189,
I take it you are a breast man - but 88D would have been more obvious. 
later
Roger...
_________________
Capitalware: Transforming tomorrow into today.
Connected to MQ!
Twitter |
|
| Back to top |
|
|
| basevij |
| Posted: Wed Oct 23, 2002 1:24 am Post subject: |
|
|
Newbie
Joined: 11 Jun 2002
Posts: 5
Location: Bristol, UK
|
Tx Guys
We fixed it by a mixture of applying the support pac (No problems there) and clearing the repositories having first killed the repos manager process.
We are all clean and healthy now and will never sully our QMs again
88D is OTT if you ask me... 
JB |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
